optimal control of a multilevel dc-link converter...
TRANSCRIPT
![Page 1: Optimal control of a multilevel DC-link converter ...eprints.whiterose.ac.uk/95668/1/Manuscript_Revised(2)_Multilevel... · 1 Optimal Control of a Multilevel DC-Link Converter Photovoltaic](https://reader031.vdocument.in/reader031/viewer/2022021817/5a9dc8ac7f8b9a85318c7dd9/html5/thumbnails/1.jpg)
This is a repository copy of Optimal control of a multilevel DC-link converter photovoltaic system for maximum power generation.
White Rose Research Online URL for this paper:http://eprints.whiterose.ac.uk/95668/
Version: Accepted Version
Article:
Abdalla, I, Corda, J and Zhang, L (2016) Optimal control of a multilevel DC-link converter photovoltaic system for maximum power generation. Renewable Energy, 92. pp. 1-11. ISSN 0960-1481
https://doi.org/10.1016/j.renene.2016.01.062
© 2016, Elsevier. Licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International http://creativecommons.org/licenses/by-nc-nd/4.0/
[email protected]://eprints.whiterose.ac.uk/
Reuse
Unless indicated otherwise, fulltext items are protected by copyright with all rights reserved. The copyright exception in section 29 of the Copyright, Designs and Patents Act 1988 allows the making of a single copy solely for the purpose of non-commercial research or private study within the limits of fair dealing. The publisher or other rights-holder may allow further reproduction and re-use of this version - refer to the White Rose Research Online record for this item. Where records identify the publisher as the copyright holder, users can verify any specific terms of use on the publisher’s website.
Takedown
If you consider content in White Rose Research Online to be in breach of UK law, please notify us by emailing [email protected] including the URL of the record and the reason for the withdrawal request.
![Page 2: Optimal control of a multilevel DC-link converter ...eprints.whiterose.ac.uk/95668/1/Manuscript_Revised(2)_Multilevel... · 1 Optimal Control of a Multilevel DC-Link Converter Photovoltaic](https://reader031.vdocument.in/reader031/viewer/2022021817/5a9dc8ac7f8b9a85318c7dd9/html5/thumbnails/2.jpg)
1
Optimal Control of a Multilevel DC-Link Converter Photovoltaic System for Maximum Power Generation
I Abdalla, J Corda, L Zhang
School of Electronic and Electrical Engineering, University of Leeds
Leeds LS2 9JT, United Kingdom E-mail: [email protected]
Abstract
This paper describes a new algorithm for optimal control of a PV system under partial shading. A
multilevel DC-link is the essential part of the proposed system and its control engages a voltage-hold
perturbation and observation (VH-P&O) method combined with a PWM algorithm with permutation
of PV sources. The algorithm enables achieving the maximum power generation for any number of
PV and converter modules. The main features of the control are: (i) a continual operation of all PV
sources, shaded and non-shaded, at their maximum power points, (ii) delivery of all extracted power
from PV sources to the load and (iii) generation of multilevel output voltage waveform with a low
total harmonic distortion.
Keywords: Multilevel Converters, MPPT, Partial Shading, Photovoltaic Panels
1 Introduction
PV power generators are commonly structured by connecting the PV modules in series to meet the
load voltage requirement. This configuration constrains that the same current flows through the chain.
However, the characteristics of the chained PV modules are not exactly identical even when they are
manufactured in the same batch. A more profound issue is the levels of illumination on individual PV
cells which may be different. Even if a single cell in the series chain is shaded, the current through the
whole chain would be reduced according to the shaded cell and consequently the overall power output
of the whole PV chain drops significantly. Furthermore the shaded cell or cells may be destroyed due
to the increased power dissipation on them resulting hot spots. A classical method to prevent high
power dissipation on shaded cells has been to connect bypass diodes across a set of chained cells
![Page 3: Optimal control of a multilevel DC-link converter ...eprints.whiterose.ac.uk/95668/1/Manuscript_Revised(2)_Multilevel... · 1 Optimal Control of a Multilevel DC-Link Converter Photovoltaic](https://reader031.vdocument.in/reader031/viewer/2022021817/5a9dc8ac7f8b9a85318c7dd9/html5/thumbnails/3.jpg)
2
(normally 18). The drawback of such a method is that under uneven irradiance, when diode/s turn on
to allow the current flowing through, the power from individual cells is lost.
Various other methods have been developed to address the problem of partial shading. One of them
uses the Module-Integrated PV Converter (MIPC) scheme [1 - 4] which is formed from a series
connection of modules comprising a PV panel and a DC/DC converter. This enables the control of
each PV source output voltage for maximum power generation, while maintaining the same current
value at the converter output terminal. However the problem with this scheme is that the converter for
the shaded PV panel may only be able to maintain the current equal to that of the other modules when
the irradiance difference is moderate.
Departure from the maximum power point (MPP) may be unavoidable for a shaded source with a
larger light level discrepancy, and in this case the option may be to control the converter to bypass the
shaded source. Another approach relies on connecting the PV sources into a number of clusters of
nearly equal extracted power by using a switching matrix to reconfigure the clusters so that the PV
sources of equal shading, hence equal power generation, are connected in parallel. The method was
presented in [5, 6] and implemented experimentally in [7]. Although tests have shown that this
approach results in more power output than in the directly connected PV sources, the cluster which
generates the least power will be bypassed completely. The third scheme, named the ‘Generation
Control Circuit’ (GCC), uses a type of multilevel DC/DC converter with multiple DC/DC converters
connected in series, each powered by a PV source [8]. The individual converter may be a buck-boost
type and can be controlled to extract power from the PV sources and deliver it to the load directly.
This idea was implemented in [9] using a fly-back converter allocated to each PV source. The main
issue with this scheme is that the PV sources cannot be controlled completely independently.
Consequently, locating the MPP of one PV source does not mean that the other PV sources operate at
their MPPs.
This paper presents a novel optimal control scheme for a PV system using a converter topology called
the multilevel DC-link inverter. Similar to the MIPC above this system consists of multiple modules
![Page 4: Optimal control of a multilevel DC-link converter ...eprints.whiterose.ac.uk/95668/1/Manuscript_Revised(2)_Multilevel... · 1 Optimal Control of a Multilevel DC-Link Converter Photovoltaic](https://reader031.vdocument.in/reader031/viewer/2022021817/5a9dc8ac7f8b9a85318c7dd9/html5/thumbnails/4.jpg)
3
connected in series, each module has a PV panel with a switch-diode pair at its output terminals. A
DC-AC bridge inverter is used at the output to convert the generated DC power into AC form [10].
The topology and control strategy enables the system to achieve the MPP tracking for all PV sources
in a system under any irradiation conditions. An initial investigation [11] has demonstrated the basic
converter topology and control concept. It has been shown that the scheme allows flexible control of
each PV source to make it operating at its respective MPP corresponding to its light level. In the
current paper a further improved control algorithm is presented which combines a voltage-hold
perturbation and observation (VH-P&O) method and the PWM algorithm with permutation of PV
sources for achieving the maximum power generation for any number of converter levels. The method
is presented by both simulation study and practical experimental test, and the results show a
significant enhancement of the PV system performance.
The paper is organised as follows. Section 2 describes the multilevel DC-link inverter topology. In
Section 3 the new control algorithm which combines the VH-P&O scheme and permutation PWM
strategy is presented in detail. Simulation study of the PV system and the control scheme, which is
performed using a STATE-SPACE AVERAGE model for the multilevel DC-link inverter, together with
the results obtained are discussed in Section 4. Finally, the experimental system and results obtained
by testing are presented in Section 5.
2 Structure of the PV System based on Multilevel DC-Link Converter
Fig. 1 shows the configuration of a PV system comprising three PV units acting in conjunction with a
multilevel DC-link converter. Each PV unit consists of a single PV source, a capacitor and a switch
with a complimentary diode. The units are connected in series and any of them can be switched in or
out of the chain by turning on or off its switch. When a unit is switched off, it is being bypassed by a
diode. The three switches (SW1, SW2, SW3) operate at a high frequency and are controlled by the
direct PWM method [12] to form a three-level positive DC voltage. The H–bridge inverter at the
output serves to convert the multilevel DC voltage waveform to alternative positive and negative
![Page 5: Optimal control of a multilevel DC-link converter ...eprints.whiterose.ac.uk/95668/1/Manuscript_Revised(2)_Multilevel... · 1 Optimal Control of a Multilevel DC-Link Converter Photovoltaic](https://reader031.vdocument.in/reader031/viewer/2022021817/5a9dc8ac7f8b9a85318c7dd9/html5/thumbnails/5.jpg)
4
output voltage half-cycles of the required output frequency (e.g. 50Hz). The generation of multiple
voltage levels, combined with proper control, enables forming the approximate sinewave output.
PV1
PV2
CPV1
CPV2
Vout
Iout
S1 S3
S2 S4
+
VPV1
-
+
VPV2
-
IPV1
IPV2
P&O
MPPT2
SW1
SW2
Vlink
sin(wrt)
SW switches operate at high
PWM frequency (kHz) S switches operate at low
frequency (50Hz)
Lo
ad
Vref1
Vref2
×
×
PV3
CPV3+
VPV3
-
IPV3
SW3
×
P&O
MPPT3
P&O
MPPT1
Vref3
PV permutation
algorithm
t
vlink
vout
Control
Unit
t
Fig. 1: PV system with multilevel DC- link converter
3 Optimal Control Scheme based on Permutation of PV Sources
The optimal control scheme should maximise the power transferred from PV sources to the AC load
or grid in different light conditions, and generate a nearly sinusoidal voltage with minimum harmonic
distortion and DC offset. Ideally, it should also ensure equal switching utilisation and hence the
losses. The solution presented in this paper comprises two parts: (i) Voltage-Hold Perturb & Observe
(VH P&O) method for generation of the MPP reference voltage and (ii ) PWM algorithm with
permutation of PV sources for switching control.
![Page 6: Optimal control of a multilevel DC-link converter ...eprints.whiterose.ac.uk/95668/1/Manuscript_Revised(2)_Multilevel... · 1 Optimal Control of a Multilevel DC-Link Converter Photovoltaic](https://reader031.vdocument.in/reader031/viewer/2022021817/5a9dc8ac7f8b9a85318c7dd9/html5/thumbnails/6.jpg)
5
3.1 The Voltage-Hold P&O Method for Reference Voltage Generation
A classical strategy for the PV maximum power point tracking is the perturbation and observation
(P&O) method, which has been widely studied and used in the PV MPPT controllers (e.g. [13-17]).
The two problems linked with this simple search method are that the terminal voltage may oscillate
around a MPP, and the system may lose the MPP during rapid irradiance changes. Some
improvements have been proposed for reducing the number of oscillations around MPP (e.g. [18,
19]), but a slow speed of the response may result in an incorrect MPPT if the changes of irradiance
are rapid.
The voltage-hold (VH) P&O method proposed here overcomes the above limitation in terms of
oscillations around MPPs, which cause noise and slow down the search process, and also it can cope
with rapid irradiance changes. For solving the problem of oscillating around a MPP, the algorithm
uses variable searching step size. As soon as the irradiation change stops, the tracking step size is
gradually decreased towards zero when reaching close to the MPP. When an irradiation change
occurs, the tracking step size will be reset to the initial value to maintain the fast tracking. At a rapid
irradiance change, which can be detected by measuring the light level and the number of changes
within a fixed time interval, the algorithm stops the searching process and sets the reference voltage to
the measured PV capacitor voltage which is essential tracking quantity. Subsequently the algorithm
resumes the perturbation process for searching the MPP voltage along the I-V characteristic
corresponding to the new light level.
The VH P&O method is applied in sequence to each PV unit at each sampling instant, to track
individually their respective reference voltages. This may lead to different voltage values depending
on the weather conditions and PV panel characteristics. For generating AC reference voltages these
values are multiplied by a unity sinusoidal signal sin のt (の = 2ヾ × 50). (For example, in the system
with three PV units, the reference voltages vref1(t), vref2(t) and vref3(t) are generated as shown in
Fig.1.)
![Page 7: Optimal control of a multilevel DC-link converter ...eprints.whiterose.ac.uk/95668/1/Manuscript_Revised(2)_Multilevel... · 1 Optimal Control of a Multilevel DC-Link Converter Photovoltaic](https://reader031.vdocument.in/reader031/viewer/2022021817/5a9dc8ac7f8b9a85318c7dd9/html5/thumbnails/7.jpg)
6
3.2 PWM Algorithm with Permutation of PV Sources
The next step should naturally be generation of PWM signals for the switches of individual PV units.
For maximum power extraction, the terminal voltage of each PV unit should be as close as possible to
its individual reference voltage established by the VH P&O process. This also implies that with
multiple PV units in a chain, the PWM reference signal for the whole system should be the sum of all
individual reference voltages, i.e. vref1(t)+ vref2(t),…+ vref(n)(t). To meet these conditions and the criteria
of equal switch utilisation and low waveform distortion, a novel PWM algorithm based on
permutation of PV sources has been developed. The algorithm consists of three stages, which are
performed within each PWM switching period Ts: the direct PWM [12], the sequential permutation of
PV sources, and the AC output voltage generation.
In Stage 1, the direct PWM determines the output voltage levels and switch-on time intervals as
follows. The reference voltages vref(j)(t) (j = 1,…n), obtained from VH P&O process, are normalised as
nV
tvv
jj /
)(
L
)ref()ref( (1)
where the standard voltage level VL equals VMPP for a single source (panel) at the standard irradiance
of 1000 W/m2 and temperature 25ºC, and n is the total number of PV sources. (This makes normalised
voltages n×vref(j)/VL, i.e. they are normalized for the full system comprising n sources.)
The integer part of each normalized reference voltage is defined as the offset voltage, i.e.
][int )ref()offset( jj vv (2)
If voffset(j) > 1, the switch of the corresponding PV source should be in 'on' state during time interval
)( )offset()ref(s)on( jjj vvTt (3)
On the other hand if voffset(j) = 0, the switch of the corresponding PV source should be in 'off' state to
allow the diode to bypass this source. Stage 1 generates all n values of voffset(j) and ton(j).
In Stage 2, the sequential permutation algorithm is implemented to determine the switching states of
all PV units for building up the output voltage of the system. All PV sources are sequentially
permutated through consecutive PWM switching cycles according to the cyclic counter C as
![Page 8: Optimal control of a multilevel DC-link converter ...eprints.whiterose.ac.uk/95668/1/Manuscript_Revised(2)_Multilevel... · 1 Optimal Control of a Multilevel DC-Link Converter Photovoltaic](https://reader031.vdocument.in/reader031/viewer/2022021817/5a9dc8ac7f8b9a85318c7dd9/html5/thumbnails/8.jpg)
7
illustrated in Fig.2. The PV source which sequentially comes in turn at the highest voltage level is the
only one which is controlled at a particular switching period (Ts), while other sources are used to build
up the appropriate voltage level. The benefits of such a process are: (i) all PV sources of the system
are equally engaged and the switching losses of transistors associated with each PV source are
balanced, (ii) the achievement of symmetrical positive and negative half-cycles of the AC output
waveform under partial shading and (iii) the reduction of the voltage ripple on capacitors of PV units.
t
Vout
Voh
Ts
PV
1
PV
2
PV
1P
V3
PV
2P
V1
PV
2P
V3
+P
V1
PV
3P
V1
+P
V2
PV
1P
V2
+P
V3
VPV
2VPV
3VPV
Vol
ton toff
C=1 C=2 C=3 C=1 C=2 C=3 C=1
Fig. 2: Permutation of PV sources through consecutive PWM switching cycles
Unlike the basic algorithm [11], the improved algorithm does not represent the PV source at the
output if its offset voltage is zero, except in the case when all offset voltages are zero. The reason for
introducing this exception is to initiate the control and it is also required in the case of low PWM
reference signals.
In Stage 3 the terminal H-bridge converter is used to change the multilevel DC voltage waveform to a
single-phase AC waveform of low frequency (e.g. 50 Hz) by tracking the sinusoidal reference signal.
Table 1 presents the switching states and output voltage generation for the system with three PV
sources (n = 3). According to formulae (1) - (2), the offset voltages voffset(j) (j = 1,2,3) can be either 0, 1
or 2. Hence there are 27 combinations in total and 27 output voltage patterns, which makes the
presentation quite cumbersome. For clarification, a graphical illustration of the process for generating
the output waveform in the basic system with two PV sources is given in the Appendix.
![Page 9: Optimal control of a multilevel DC-link converter ...eprints.whiterose.ac.uk/95668/1/Manuscript_Revised(2)_Multilevel... · 1 Optimal Control of a Multilevel DC-Link Converter Photovoltaic](https://reader031.vdocument.in/reader031/viewer/2022021817/5a9dc8ac7f8b9a85318c7dd9/html5/thumbnails/9.jpg)
8
Table 1: Switching states and output voltage generation for the system with three PV sources
C voffset1 voffset2 voffset3 Vol Voh ton
1 0 0 0 0 VPV1 ton1
0 0 1 0 0
0 0 2 VPV3 VPV3
0 1 0 VPV2 VPV2
0 1 1 VPV2 VPV2
0 1 2 VPV2+VPV3 VPV2+VPV3
0 2 0 VPV2 VPV2
0 2 1 VPV2+VPV3 VPV2+VPV3
0 2 2 VPV2+VPV3 VPV2+VPV3
1 0 0 0 VPV1
1 0 1 0 VPV1
1 0 2 VPV3 VPV1+ VPV3
1 1 0 VPV2 VPV1+ VPV2
1 1 1 VPV2 VPV1+ VPV2
1 1 2 VPV2+VPV3 VPV1+ VPV2+ VPV3
1 2 0 VPV2 VPV1+ VPV2
1 2 1 VPV2+VPV3 VPV1+ VPV2+ VPV3
1 2 2 VPV2+VPV3 VPV1+ VPV2+ VPV3
2 0 0 0 VPV1
2 0 1 VPV3 VPV1+ VPV3
2 0 2 VPV3 VPV1+ VPV3
2 1 0 VPV2 VPV1+ VPV2
2 1 1 VPV2+VPV3 VPV1+ VPV2+ VPV3
2 1 2 VPV2+VPV3 VPV1+ VPV2+ VPV3
2 2 0 VPV2 VPV1+ VPV2
2 2 1 VPV2+VPV3 VPV1+ VPV2+ VPV3
2 2 2 VPV2+VPV3 VPV1+ VPV2+ VPV3
2 0 0 0 0 VPV2 ton2
0 0 1 VPV3 VPV3
0 0 2 VPV3 VPV3
0 1 0 0 VPV2
0 1 1 VPV3 VPV2+VPV3
0 1 2 VPV3 VPV2+VPV3
0 2 0 0 VPV2
0 2 1 VPV3 VPV2+VPV3
0 2 2 VPV3 VPV2+VPV3
1 0 0 0 0
1 0 1 VPV3 VPV3
1 0 2 VPV1+VPV3 VPV1+VPV3
1 1 0 0 VPV2
1 1 1 VPV3 VPV2+ VPV3
1 1 2 VPV1+VPV3 VPV1+ VPV2+ VPV3
1 2 0 VPV1 VPV1+ VPV2
1 2 1 VPV1+VPV3 VPV1+ VPV2+ VPV3
1 2 2 VPV1+VPV3 VPV1+ VPV2+ VPV3
2 0 0 VPV1 VPV1
2 0 1 VPV1+VPV3 VPV1+ VPV3
2 0 2 VPV1+VPV3 VPV1+ VPV3
2 1 0 VPV1 VPV1+ VPV2
2 1 1 VPV1+VPV3 VPV1+ VPV2+ VPV3
2 1 2 VPV1+VPV3 VPV1+ VPV2+ VPV3
2 2 0 VPV1 VPV1+VPV2
2 2 1 VPV1+VPV3 VPV1+ VPV2+ VPV3
2 2 2 VPV1+VPV3 VPV1+ VPV2+ VPV3 c
Continued …
C voffset1 voffset2 voffset3 Vol Voh ton
3 0 0 0 0 VPV3 ton3
0 0 1 0 VPV3
0 0 2 0 VPV3
0 1 0 0 0
0 1 1 0 VPV3
0 1 2 VPV2 VPV2+VPV3
0 2 0 VPV2 VPV2
0 2 1 VPV2 VPV2+VPV3
0 2 2 VPV2 VPV2+VPV3
1 0 0 VPV1 VPV1
1 0 1 VPV1 VPV1+ VPV3
1 0 2 VPV1 VPV1+ VPV3
1 1 0 VPV1 VPV1
1 1 1 VPV1 VPV1+ VPV3
1 1 2 VPV1+VPV2 VPV1+ VPV2+ VPV3
1 2 0 VPV1+VPV2 VPV1+ VPV2
1 2 1 VPV1+VPV2 VPV1+ VPV2+ VPV3
1 2 2 VPV1+VPV2 VPV1+ VPV2+ VPV3
2 0 0 VPV1 VPV1
2 0 1 VPV1 VPV1+ VPV3
2 0 2 VPV1 VPV1+ VPV3
2 1 0 VPV1+VPV2 VPV1+ VPV2
2 1 1 VPV1+VPV2 VPV1+ VPV2+ VPV3
2 1 2 VPV1+VPV2 VPV1+ VPV2+ VPV3
2 2 0 VPV1+VPV2 VPV1+ VPV2
2 2 1 VPV1+VPV2 VPV1+ VPV2+ VPV3
2 2 2 VPV1+VPV2 VPV1+ VPV2+ VPV3
![Page 10: Optimal control of a multilevel DC-link converter ...eprints.whiterose.ac.uk/95668/1/Manuscript_Revised(2)_Multilevel... · 1 Optimal Control of a Multilevel DC-Link Converter Photovoltaic](https://reader031.vdocument.in/reader031/viewer/2022021817/5a9dc8ac7f8b9a85318c7dd9/html5/thumbnails/10.jpg)
9
4 Simulation of the PV System Performance
The multilevel DC-link converter is simulated using a state-space average (SSA) model having the
general form given by [20]:
)pv(
2pv
1pv
)pv(
2pv
1pv
)pv(
2pv
1pv
)pv(
2
)pv(
2
)pv(
1
2pv
2
2pv
22
2
12
1pv
1
1pv
21
1pv
21
)pv(
2pv
1pv
100000
00000
0000
000
001
0
001
1
d
d
d
dd
d
n
n
n
n
n
n
n
n
n
n
pv
n
ni
i
i
C
C
C
v
v
v
C
D
C
DD
C
DD
C
DD
C
D
C
DD
C
DD
C
DD
C
D
R
t
v
t
vt
v
(4)
)pv(
2pv
1pv
21out )12(
n
n
v
v
v
DDDDv
(5)
where, the input vector elements are the currents ipv1, ipv2,…ipv(n) of PV sources; the terminal voltages
vpv1, vpv2,…vpv(n) form the state vector; Cpv1, Cpv2,…Cpv(n) are the corresponding terminal capacitances;
Dl, D2, … Dn are the state variables for switches SW1, SW2… SW(n); R is the load resistance; D is the
state variable of the single-phase output H-bridge inverter (D = 1 when S1&S4 are switched on, and D
= 0 when S2&S3 are switched on).
Elements of the input vector [ipv1, ipv2,…ipv(n)] are found from the MPPs of the I-V characteristics
which are generated using Bishop’s model for the PV source [21]. The I-V characteristics for different
irradiance levels are shown in Fig.3 with indicated MPPs. Table 2 lists the system parameters used in
the model.
![Page 11: Optimal control of a multilevel DC-link converter ...eprints.whiterose.ac.uk/95668/1/Manuscript_Revised(2)_Multilevel... · 1 Optimal Control of a Multilevel DC-Link Converter Photovoltaic](https://reader031.vdocument.in/reader031/viewer/2022021817/5a9dc8ac7f8b9a85318c7dd9/html5/thumbnails/11.jpg)
10
Voltage (V) 0
5
10
15
20
0 5 10 15 20 25 30 35 40 45 50 55 60 64
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
5
5.5
6
Ipv(
A)
2
21000 W/m
2750 W/m
2500 W/m
250 W/m
5 A, 50 V250 W
3.74 A, 48.62 V181.9 W
2.48 A, 46.66 V115.9 W
1.23 A, 43.33 V53.38 W
Cu
rre
nt (A
)
Fig. 3: I-V characteristics of the PV source for different irradiance levels
(MPP is indicated on each characteristic.)
Table 2: PV system parameters used in simulations
The waveforms were simulated for the PV systems with n = 2, 3, 4 and 5 sources generating
respectively five, seven, nine and eleven-level AC output voltages. The Fast Fourier Transformation
(FFT) is applied to the output waveform to evaluate the fundamental harmonic and the total
harmonics distortion (THD).
Symbol Parameter Value
PMPP Maximum power of PV source 250 W
Cpv PV source terminal capacitor 5500 ȝF
R Load resistance
8.5 Ω (5&7-level)
18.5 Ω (9-level)
23 Ω (11-level)
ev P&O tracking step size 0.5 V
Mf Frequency modulation index 100 (5-level)
300 (7,9&11-level)
f AC output frequency 50 Hz
![Page 12: Optimal control of a multilevel DC-link converter ...eprints.whiterose.ac.uk/95668/1/Manuscript_Revised(2)_Multilevel... · 1 Optimal Control of a Multilevel DC-Link Converter Photovoltaic](https://reader031.vdocument.in/reader031/viewer/2022021817/5a9dc8ac7f8b9a85318c7dd9/html5/thumbnails/12.jpg)
11
Fig. 4a shows five-level output voltage waveforms produced by the system with two sources PV1 and
PV2 exposed to irradiances of 500 W/m2 and 1000 W/m2 respectively. The improved algorithm gives
better voltage waveforms (shown on the right) compared with its predecessor. The THD of 43.5% and
the 50-Hz fundamental peak voltage V1(p) = 76.8 V were achieved with the improved algorithm,
compared to the THD of 58.65% and V1(p) = 70.7 V achieved with the predecessor. Similar comparison
was made for the seven-level converter with three PV sources at irradiances of 1000, 500 and 1000
W/m2 applied to PV1, PV2 and PV3 respectively. The output voltage waveforms are shown in Fig.
4b. The improved algorithm generates substantially better waveform with THD = 35.72% and V1(p) =
108.64 V, compared to the THD = 57.41% and V1(p) = 95.81 V achieved with the predecessor.
Fig. 4: Output voltage waveforms simulated by the predecessor and the improved algorithm.
(a) five-level AC waveforms with sources PV1 and PV2 at irradiances 500 and 1000 W/m2;
(b) seven-level AC waveforms with sources PV1, PV2 and PV3 at irradiances 1000, 500 and
1000 W/m2 respectively.
Predecessor algorithm Improved permutation algorithm
1.8 1.805 1.81 1.815 1.82-100-75-50-25
0255075
100
Vout
(V)
1.35 1.355 1.36 1.365 1.37-100-75-50-25
0255075
100THD=58.65%,V1(p)=70.72V
THD=43.52%,V1(p)=76.81V
Time (s) Time (s)
(a)
Predecessor algorithm Improved permutation algorithm
3.27 3.275 3.28 3.285 3.29-150
-100
-50
0
50
100
150
Vou
t (V
)
4.99 4.995 5 5.005 5.01-150
-100
-50
0
50
100
150THD=57.41%,V1(p)=95.81V
THD=35.72%,V1(p)=108.64V
Time (s) Time (s)
(b)
![Page 13: Optimal control of a multilevel DC-link converter ...eprints.whiterose.ac.uk/95668/1/Manuscript_Revised(2)_Multilevel... · 1 Optimal Control of a Multilevel DC-Link Converter Photovoltaic](https://reader031.vdocument.in/reader031/viewer/2022021817/5a9dc8ac7f8b9a85318c7dd9/html5/thumbnails/13.jpg)
12
Fig. 5 illustrates output voltage waveforms and power extracted from the system with four PV sources
using 9-level converter for two cases – (a) none of the sources are shaded and (b) one source is
shaded. The power levels extracted from individual sources and the load power are shown by line
graphs on the right-hand side. Referring to Fig. 5a, when all four PV sources are exposed to equal
irradiance of 1000 W/m2, the THD is 18.4%, the 50-Hz fundamental peak voltage V1(p) = 194 V and
the power extracted from each panel is 250 W giving 1000 W to the load. Referring to Fig. 5b, when
one source is shaded, i.e. exposed to the halved irradiance (500 W/m2), the predicted THD is 21.67%,
V1(p) = 176 V and the total power delivered to the load is 865 W conforming that it is very close to
the sum of maximum power values found from I-V characteristics (3x250+116 W).
Fig. 5: Simulated output voltage waveforms and the extracted power achieved by the improved
algorithm for 9-level converter with four PV sources.
(a) all at the same irradiance of 1000W/m2;
(b) at irradiances of 1000 W/m2 for PV1, PV2, PV3 and 500W/m2 for PV4.
2.81 2.815 2.82 2.825 2.83
-200
-100
0
100
200
Vou
t (V
)
2 2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9 30
100200300400500600700800900
1,000
Pow
er (W
)
PV1,2,3,4
Load pow erTHD=18.40%
V1(p)=194.16V
Time (s) Time (s)
(a)
2.55 2.555 2.56 2.565 2.57
-200
-100
0
100
200
Vou
t (V
)
2 2.2 2.4 2.6 2.8 30
100200300400500600700800900
1,000
Pow
er (w
)
PV1,2,3
Load pow er
PV4
THD=21.67%V1(p)=176.26V
Time (s) Time (s)
(b)
![Page 14: Optimal control of a multilevel DC-link converter ...eprints.whiterose.ac.uk/95668/1/Manuscript_Revised(2)_Multilevel... · 1 Optimal Control of a Multilevel DC-Link Converter Photovoltaic](https://reader031.vdocument.in/reader031/viewer/2022021817/5a9dc8ac7f8b9a85318c7dd9/html5/thumbnails/14.jpg)
13
Fig. 6 illustrates another example, which relates to the system with five PV sources and using 11-level
converter. Referring to Fig. 6a, when all sources are exposed to equal irradiance of 1000 W/m2, the
values of THD and the 50-Hz fundamental peak voltage V1(p) are respectively 14.95% and 241 V and
both are improved compared to the values achieved by 9-level converter (18.4% and 194 V). Fig. 6b
relates to the state when PV sources are exposed to different irradiances (250, 500, 750, 750 and 1000
W/m2). The predicted THD is now 19.48%, and the total power delivered to the load is 790W which
is very close to the sum of maximum power values of five PV sources found from I-V characteristics
(53W, 116 W, 2x182 W and 250 W).
Fig. 6: Simulated output voltage waveforms and the extracted power achieved by the improved
algorithm for 11-level converter with five PV sources.
(a) all at the same irradiance of 1000W/m2;
(b) at irradiances of 250, 500, 750, 750 and 1000 W/m2.
Fig. 7 illustrates the dynamic performance in terms of extracted power from each PV source and the
total power delivered to the load from 7-level converter for a sequence of irradiance variations.
Initially the full irradiance of 1000 W/m2 is applied to all three PV sources and 250 W is extracted
3.71 3.715 3.72 3.725 3.73
-200
-100
0
100
200
Vou
t (V
)
2.5 2.6 2.7 2.8 2.9 30
100200300400500600700800900
1,0001,1001,2001,300
Pow
er (W
)
PV1,2,3,4,5
Load power
THD=14.95%V1(p)=240.67V
Time (s) Time (s)
(a)
3.85 3.855 3.86 3.865 3.87
-200
-100
0
100
200
Vou
t (V
)
2.5 2.6 2.7 2.8 2.9 30
100200300400500600700800900
1,0001,1001,2001,300
Pow
er
(W)
PV3,4
Load power
THD=19.48%V1(p)=209.12V
PV5PV2PV1
Time (s) Time (s)
(b)
![Page 15: Optimal control of a multilevel DC-link converter ...eprints.whiterose.ac.uk/95668/1/Manuscript_Revised(2)_Multilevel... · 1 Optimal Control of a Multilevel DC-Link Converter Photovoltaic](https://reader031.vdocument.in/reader031/viewer/2022021817/5a9dc8ac7f8b9a85318c7dd9/html5/thumbnails/15.jpg)
14
from each source delivering a total of 750 W to the load. In the time between 2 and 5 seconds, the
irradiance to PV3 source is halved (500 W/m2), while PV1 and PV2 are still at full irradiance. The
extracted power from PV3 is now reduced to 116 W, which matches the MPP on the I-V
characteristic at irradiance of 500 W/m2 in Fig.3. The power levels from PV1 and PV2 are not
affected by the shading of PV3, so the total delivered power is 616 W. During the time interval
between 5 and 8 seconds, both PV2 and PV3 are shaded at the halved irradiance (500 W/m2) while
PV1 is under full irradiance, and the delivered power is 482 W. Between 8 and 11 seconds, all three
PV sources are shaded to the halved irradiance (500W/m2), and the power delivered to the load is
reduced to 348 W. After 11 seconds all three PV sources are again exposed to the full irradiance and
are recovered to their MPPs of 250 W delivering the power of 750 W to the load.
Fig. 7. Power responses of the system with seven-level DC-link converter during irradiance changes.
(a) irradiance variations in time;
(b)-(d) extracted power from individual sources PV1, PV2 and PV3;
(e) total power delivered to the load.
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7 7.5 8 8.5 9 9.5 1010.51111.51270
100130160190220250
PV
1 P
ower
(W
)
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7 7.5 8 8.5 9 9.5 1010.51111.51270
100130160190220250
PV
2 P
ower
(W
)
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7 7.5 8 8.5 9 9.5 1010.51111.512100125150175200225250
PV
3 P
ower
(W
)
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7 7.5 8 8.5 9 9.5 1010.51111.512250350450550650750
Load
pow
er (
W)
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7 7.5 8 8.5 9 9.5 1010.51111.512450550650750850950
1050
Irra
dian
ce (
W/m
)
Time (s)
PV3 PV2 PV1
2
![Page 16: Optimal control of a multilevel DC-link converter ...eprints.whiterose.ac.uk/95668/1/Manuscript_Revised(2)_Multilevel... · 1 Optimal Control of a Multilevel DC-Link Converter Photovoltaic](https://reader031.vdocument.in/reader031/viewer/2022021817/5a9dc8ac7f8b9a85318c7dd9/html5/thumbnails/16.jpg)
15
5 Experimental Results
Laboratory experiments were performed on a system having three identical small serially connected
PV modules (sources) with an in-house built 7-level converter. Each module was exposed to a
separate artificial sunlight with adjustable irradiance emulated by three halogen bulbs each of which
was supplied from a different phase of a 3-phase variac connected to the 3-phase 50-Hz mains supply.
This arrangement provides a fairly uniform irradiance to each module.
The parameters of the experimental system are listed in Table 3. The measured I-V characteristic of
the PV module are shown in Fig. 8 under different irradiance levels at room and surface temperatures
of 20°C and 38°C respectively.
Table 3: Experimental system parameters
Though the power ratings of the PV panels are lower than those used in the simulation study, this is
considered adequate for verifying the principle of the proposed control scheme for the following
reasons: (i) With the artificial sun light simulator set up in the available laboratory, the measured I-V
characteristics of small PV panels corresponding to various light and temperature levels follow
closely the characteristics given in the data sheets. However, panels with higher power rating (over
100 W) give significantly lower output power compared to what they should generate based on the
data sheet; (ii) For validating the proposed control scheme the difference in power rating is not
considered an issue, because the controller can be scaled up for higher power system without causing
dynamic stability problems.
Symbol Quantity Value
PMPP PV source maximum power 10 W
Cpv PV source terminal capacitor 2200 ȝF
R Load resistance 30 Ω
ev P&O tracking step size 0.2 V
Mf Frequency modulation index 100
f Output frequency 50 Hz
![Page 17: Optimal control of a multilevel DC-link converter ...eprints.whiterose.ac.uk/95668/1/Manuscript_Revised(2)_Multilevel... · 1 Optimal Control of a Multilevel DC-Link Converter Photovoltaic](https://reader031.vdocument.in/reader031/viewer/2022021817/5a9dc8ac7f8b9a85318c7dd9/html5/thumbnails/17.jpg)
16
Fig. 8. Measured I-V characteristic of a 10W PV module at different irradiances
The improved control algorithm has been applied to the experimental system. The control is
implemented via a digital processing unit eZdspTM F28335 in real-time [22], and the Matlab-Simulink
real time data exchange (RTDX) is employed to display the extracted power and to modify the system
parameters via a graphical user interface (GUI). Details of this are presented in [21].
The extracted power from each PV module can be monitored on the bar chart of the GUI window in
real-time mode as illustrated in Fig. 9 which shows two examples with different shading conditions.
The graphs demonstrate that the developed system ensures that shaded PV module/s do not affect the
operation of non-shaded module/s which continue to generate their maximum power.
Fig. 9: GUI window showing the extracted power from each PV module at different irradiances
1000 W/m2
725 W/m2
580 W/m2
435 W/m2
290 W/m2
145 W/m2
70 W/m2
0
0.1
0.2
0.3
0.4
0.5
0.6
Voltage (V) 0
5
10
15
20
0.7
Cur
rent
(A)
![Page 18: Optimal control of a multilevel DC-link converter ...eprints.whiterose.ac.uk/95668/1/Manuscript_Revised(2)_Multilevel... · 1 Optimal Control of a Multilevel DC-Link Converter Photovoltaic](https://reader031.vdocument.in/reader031/viewer/2022021817/5a9dc8ac7f8b9a85318c7dd9/html5/thumbnails/18.jpg)
17
Table 4 summarizes the THD and the 50Hz fundamental amplitude of the load voltage waveforms
which are shown in Fig. 10 for both the previous and improved PV permutation algorithms as a
function of the PV extracted power. It is apparent that at the same conditions of partial shading, the
improved PV permutation algorithm has resulted in a lower output harmonic distortion and larger
amplitude of the fundamental harmonic (50 Hz), compared to the precursor algorithm.
Fig.10: Output voltage and current waveforms of 7-level converter: measured under control by
precursor algorithm (column 1); measured and simulated under control by improved
algorithm (columns 2 and 3). Irradiances applied to sources PV1, PV2 and PV3 in W/m2,
top to bottom row: (1000, 500, 1000); (1000, 1000, 150); (1000, 750, 250); (500, 1000, 250).
-4
-3
-2
-1
0
1
2
3
4
-4
-3
-2
-1
0
1
2
3
4
-4
-3
-2
-1
0
1
2
3
4
-4
-3
-2
-1
0
1
2
3
4
THD=40.26%, V1(p)=35.16V
THD=34.80%, V1(p)=33.26V THD=55.21%, V1(p)=30.83V
THD=40.50%, V1(p)=31.37V THD=64.92%, V1(p)=29.91V
THD=38.46%, V1(p)=29.07V THD=61.22%, V1(p)=27.95V THD=43.65%, V1(p)=29.37V
THD=42.38%, V1(p)=31.80V
Vol
tage
(V)
60 40 20 0 -20 -40 -60
Cur
rent
(A) 2
0 -2
Vol
tage
(V)
60 40 20 0 -20 -40 -60
Cur
rent
(A) 2
0 -2
10ms
Vol
tage
(V)
60 40 20 0 -20 -40 -60
Cur
rent
(A) 2
0 -2
Vol
tage
(V)
60 40 20 0 -20 -40 -60
Cur
rent
(A) 2
0 -2
THD=36.00%, V1(p)=36.33V THD=34.05%, V1(p)=36.23V
THD=33.68%, V1(p)=34.18V
(a)
(b)
(c)
(d)
iout
vout
iout
vout
iout
vout
Time (10 ms / div) Time (10 ms / div) Time (10 ms / div)
![Page 19: Optimal control of a multilevel DC-link converter ...eprints.whiterose.ac.uk/95668/1/Manuscript_Revised(2)_Multilevel... · 1 Optimal Control of a Multilevel DC-Link Converter Photovoltaic](https://reader031.vdocument.in/reader031/viewer/2022021817/5a9dc8ac7f8b9a85318c7dd9/html5/thumbnails/19.jpg)
18
Table 5 lists the numerical values obtained by measurements and simulation of the power delivered to
the load when the improved control algorithm is applied. The small discrepancies between measured
and simulated results are principally caused by the inverter losses which were not accounted for in
simulations.
Table 4: Summary of measured and simulated results at various shading levels
Table 5: Measured and simulated values of power delivered to the load at various shading levels
6 Conclusions
The main features of the newly developed control scheme, which uses the VH P&O method and
improved PV source PWM permutation algorithm, are: (i) the upholding of operation of all PV
sources (shaded or not shaded) at their MPPs, (ii) the delivery of all extracted power from PV sources
to the load and (iii) the generation of improved multilevel voltage waveforms with low THD.
Extracted power (W) Precursor algorithm
Measured results Improved algorithm
Measured results Improved algorithm Simulated Results
PV1 PV2 PV3 THD (%) V1(P) (V) THD (%) V1(P) (V) THD
(%) V1(P) (V)
10 5 10 40.26 35.16 36.00 36.33 34.05 36.23
10 10 1.5 55.21 30.83 34.80 33.26 33.68 34.18
10 7.5 2.5 64.92 29.91 40.50 31.37 42.38 31.80
5 10 2.5 61.22 27.95 38.46 29.07 43.65 29.37
Extracted power (W) Measured
Load Power (W)
Simulated
Load Power (W) PV1 PV2 PV3
10 10 10 28.89 29.96
10 5 10 23.81 24.96
10 10 1.5 21.01 21.44
10 7.5 2.5 19.34 19.98
5 10 2.5 17.20 17.49
![Page 20: Optimal control of a multilevel DC-link converter ...eprints.whiterose.ac.uk/95668/1/Manuscript_Revised(2)_Multilevel... · 1 Optimal Control of a Multilevel DC-Link Converter Photovoltaic](https://reader031.vdocument.in/reader031/viewer/2022021817/5a9dc8ac7f8b9a85318c7dd9/html5/thumbnails/20.jpg)
19
The proposed system was simulated using Matlab–Simulink for multilevel DC link converters with
five, seven, nine and eleven-level AC output voltages. Static and dynamic irradiance levels were
applied to test the system behaviour with respect to the MPP tracking and the quality of the output
waveforms. The comparison between the previous and the improved algorithm have shown that the
latter generates a lower distortion of the waveform and higher amplitude of the fundamental (50-Hz)
harmonic of the output voltage. Experimental tests on a small-scale system, which uses three PV
sources and seven-level DC-link converter, have validated the advantages of the new scheme.
The time varying irradiance levels were applied to test dynamic behaviour of the system in terms of
the MPP tracking. The algorithm enables the recovery of each shaded PV source to its MPP without
affecting other PV sources connected in the chain.
Appendix
When C = 1:
If voffset1 is either 0 or 1 and voffset2 = 0, then:
SW1 is ‘on’ during interval ton1 and ‘off’ otherwise;
SW2 is ‘off’ during entire period Ts.
If voffset1 = 0 and voffset2 = 1, then:
SW1 is ‘off’ and SW2 is ‘on’ during entire period Ts.
If voffset1 = 1 and voffset2 = 1, then:
SW1 is ‘on’ during interval ton1 and ‘off’ otherwise;
SW2 is ‘on’ during entire period Ts.
When C = 2:
If voffset2 is either 0 or 1 and voffset1 = 0, then:
SW1 is ‘off’ during entire period Ts;
SW2 is ‘on’ during interval ton2 and ‘off’ otherwise.
If voffset2 is 0 and voffset1 = 1, then:
SW2 is ‘off’ and SW1 is ‘on’ during entire period Ts.
If voffset1 = 1 and voffset2 = 1, then:
SW1 is ‘on’ during entire period Ts;
SW2 is ‘on’ during interval ton2 and ‘off’ otherwise.
![Page 21: Optimal control of a multilevel DC-link converter ...eprints.whiterose.ac.uk/95668/1/Manuscript_Revised(2)_Multilevel... · 1 Optimal Control of a Multilevel DC-Link Converter Photovoltaic](https://reader031.vdocument.in/reader031/viewer/2022021817/5a9dc8ac7f8b9a85318c7dd9/html5/thumbnails/21.jpg)
20
In Fig. A1, vref1 and vref2 are two reference voltages of the five-level converter. These signals are
sampled at periods Ts yielding Mf samples for each reference signal over the time period Tr = 1/fr. (fr
denotes the required output frequency, e.g. 50 Hz.) At each period Ts, the sampled values are applied
to the Eqs. (1) - (3) for generating the required control signals.
t
vref2
t
vout
TS
VL
t
t
t
t
SW1
SW2
S1S4
S2S3
t
vref1
VL
1C 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1
VP
V2
VP
V2
VP
V1
VP
V2
VP
V1
VP
V2
VP
V1
VP
V2
VP
V1
VP
V2
VP
V1
VP
V1
VP
V1
VP
V2 +
VP
V1
VP
V1 +
VP
V2
TS Tr /2
Fig. A1: Graphical illustration of the process for generating the output waveform in the basic
system with two PV sources
![Page 22: Optimal control of a multilevel DC-link converter ...eprints.whiterose.ac.uk/95668/1/Manuscript_Revised(2)_Multilevel... · 1 Optimal Control of a Multilevel DC-Link Converter Photovoltaic](https://reader031.vdocument.in/reader031/viewer/2022021817/5a9dc8ac7f8b9a85318c7dd9/html5/thumbnails/22.jpg)
21
References
[1] G. R. Walker, P. C. Sernia, "Cascaded DC-DC converter connection of photovoltaic modules",
IEEE Trans. Power Electronics, vol.19, no.4, pp.1130-1139, July 2004.
[2] E. Roman, R. Alonso, P. Ibanez, S. Elorduizapatarietxe, D. Goitia, "Intelligent PV Module for Grid-Connected PV Systems", IEEE Trans. Industrial Electronics, vol.53, no.4, pp.1066-1073, June 2006.
[3] B. Chong, L. Zhang, “Controller Design for Integrated PV-Converter Modules under Partial Shading Conditions”, ELSEVIER Journal of International Solar Energy Society, vol.92, pp.123-138, 2013.
[4] S. Busquets-Monge, J. Rocabert, P. Rodriguez, S. Alepuz, J. Bordonau, "Multilevel Diode-Clamped Converter for Photovoltaic Generators With Independent Voltage Control of Each Solar Array", IEEE Trans. Industrial Electronics, vol.55, no.7, pp.2713-2723, July 2008.
[5] L. Gao, R. A. Dougal, S. Liu, A. Iotova, “Parallel-connected solar PV system to address partial and rapidly fluctuating shadow conditions”, IEEE Trans. Ind. Electron., vol. 56, no. 5, pp. 1548–1556, May 2009.
[6] G. Velasco, J. J. Negroni, F. Guinjoan, R. Pique, "Energy generation in PV grid-connected systems: power extraction optimization for plant oriented PV generators", IEEE International Symposium on Industrial Electronics (ISIE), vol.3, pp.1025-1030, 2005.
[7] G. Velasco-Quesada, F. Guinjoan-Gispert, R. Pique-Lopez, M. Roman-Lumbreras, A. Conesa-Roca, "Electrical PV Array Reconfiguration Strategy for Energy Extraction Improvement in Grid-Connected PV Systems", IEEE Trans. Industrial Electronics, vol.56, no.11, pp.4319-4331, Nov. 2009.
[8] T. Shimizu, M. Hirakata, T. Kamezawa, H. Watanabe, "Generation control circuit for photovoltaic modules", IEEE Trans. Power Electronics, vol.16, no.3, pp.293-300, May 2001.
[9] Qi Zhang, Xiangdong Sun, Yanru Zhong, Mikihiko Matsui, "A novel topology for solving the partial shading problem in photovoltaic power generation system", IEEE 6th International Power Electronics and Motion Control Conf. (IPEMC '09), pp.2130-2135, 2009.
[10] S. J. Lee, H. S. Bae, B. H. Cho, "Modeling and control of the single-phase photovoltaic grid-connected cascaded H-bridge multilevel inverter", IEEE Energy Conversion Congress and Exposition Conf. (ECCE’09), pp.43-47, 2009.
[11] I. Abdalla, J. Corda, L. Zhang, "Multilevel DC-Link Inverter and Control Algorithm to Overcome the PV Partial Shading," IEEE Trans. Power Electronics, vol.28, no.1, pp.14-18, Jan. 2013.
[12] D. Ning-Yi, W. Man-Chung, C. Yuan-Hua, H. Ying-Duo, "A 3-D generalized direct PWM algorithm for multilevel converters", IEEE Letters. Power Electronics, vol.3, no.3, pp.85-88, 2005.
![Page 23: Optimal control of a multilevel DC-link converter ...eprints.whiterose.ac.uk/95668/1/Manuscript_Revised(2)_Multilevel... · 1 Optimal Control of a Multilevel DC-Link Converter Photovoltaic](https://reader031.vdocument.in/reader031/viewer/2022021817/5a9dc8ac7f8b9a85318c7dd9/html5/thumbnails/23.jpg)
22
[13] J. H. R. Enslin, "Maximum Power Point Tracking: A Cost Saving Necessity in Solar Energy Systems", 16th Annual Conference of IEEE Industrial Electronics Society (IECON '90), pp. 1073-1077, 1990.
[14] K. H. Hussein, I. Muta, T. Hshino, and M. Osakada, “Maximum photovoltaic power tracking: an algorithm for rapidly changing atmospheric conditions,” IEE Procceedings – Generation, Transmission and Distribution, vol. 142, no.1, pp. 59-64, 1995.
[15] C. Hua, J Lin and C. Shen, "Implementation of a DSP-Controlled Photovoltaic System with Peak Power Tracking", IEEE Transactions on Industrial Electronics, vol.45, no.1, pp. 99-107, 1998.
[16] M. Matsui, T. Kitano, D. Xu and Z. Yang "A New Maximum Photovoltaic Power Tracking Control Scheme based on Power Equilibrium at DC Link", Conference Record of the IEEE Thirty-Fourth IAS Annual Meeting, vol.2, pp. 804-809, 1999.
[17] N. Kasa, T. Iida and H. Iwamoto, "Maximum Power Point Tracking with Capacitor Identificator for Photovoltaic Power System", Eighth International Conference on Power Electronics and Variable Speed Drives, IEE Conf. Publ. No.475, pp. 130-135, 2000.
[18] E. Koutroulis, K. Kalaitzakis and N. C. Voulgaris, "Development of a Microcontroller-based Photovoltaic Maximum Power Point Tracking Control System", IEEE Trans. on Power Electronics, vol.16, no.1, pp.46-54, 2001.
[19] M. Veerachary, T. Senjyu, and K. Uezato, "Voltage-based maximum power point tracking control of PV system", IEEE Trans. Aerosp. Electron. Syst., vol.38, no.1, pp.262-270, 2002.
[20] R. D. Middlebrook and S. Cuk, "A generalized unified approach to modelling switching-converter power stages", IEEE Power Electronics Specialist Conference, PESC'76 Record, pp.18-34, 1976.
[21] I. Abdalla, “Integrated PV and multilevel converter system for maximum power generation under partial shading conditions”, PhD Thesis, University of Leeds, UK, 2013.
[22] Texas instruments, TMS320F28335 Data Manual, 2010.